How to integrate RudderStack with a Gatsby website

8 minute read

You can quickly and easily get RudderStack up and running in your Gatsby application by using RudderStack’s Gatsby plugin. It makes it easy to integrate your Gatsby website with the RudderStack API and easily track events.

Key features of this plugin

Use multiple write keys (one for production environment, another optional one for development)
Disable page view tracking (just in case you want to add it later manually)
Always up-to-date and maintained by the RudderStack team

Installing the plugin

To install the plugin, run the following command with the package manager of your choice:

NPM: $ npm install --save gatsby-plugin-rudderstack
YARN: $ yarn add gatsby-plugin-rudderstack

Setup

Step 1: Configure your Gatsby config file

In your gatsby-config.js file you first need to add your write key from your account. You can find this in your RudderStack dashboard.

Configuration options

plugins: [
  {
    resolve: `gatsby-plugin-rudderstack`,
    options: {
      // your rudderstack write key for your production environment
      // when process.env.NODE_ENV === 'production'
      // required; non-empty string
      //NOTE: Do not commit this to git. Process from env.
      prodKey: `RUDDERSTACK_PRODUCTION_WRITE_KEY`,
      // if you have a development env for your rudderstack account, paste that key here
      // when process.env.NODE_ENV === 'development'
      // optional; non-empty string
      //NOTE: Do not commit this to git. Process from env.
      devKey: `RUDDERSTACK_DEV_WRITE_KEY`,
      // boolean (defaults to false) on whether you want
      // to include analytics.page() automatically
      // if false, see below on how to track pageviews manually
      trackPage: false,
      // number (defaults to 50); time to wait after a route update before it should
      // track the page change, to implement this, make sure your `trackPage` property is set to `true`
      trackPageDelay: 50,
      // If you need to proxy events through a custom data plane,
      // add a `dataPlaneUrl` property (defaults to https://hosted.rudderlabs.com )
      // RudderStack docs:
      //   - https://rudderstack.com/docs/sources/event-streams/sdks/rudderstack-javascript-sdk/#31-load
      dataPlaneUrl: `https://override-rudderstack-endpoint`,
      // Add a `controlPlaneUrl` property if you are self-hosting the Control Plane
      // RudderStack docs:
      //   - https://rudderstack.com/docs/sources/event-streams/sdks/rudderstack-javascript-sdk/#311-self-hosted-control-plane
      controlPlaneUrl: `https://override-control-plane-url`,
      // boolean (defaults to false); whether to delay loading (add SDK script and load the SDK) of RudderStack JS SDK.
      // This feature will force RudderStack to load _after_ either a page routing change
      // or user scroll, whichever comes first. This delay time is controlled by
      // `delayLoadTime` setting. This feature is used to help improve your website's
      // TTI (for SEO, UX, etc).  See links below for more info.
      // NOTE: But if you are using server-side routing and enable this feature,
      // RudderStack will never load (because although client-side routing does not do
      // a full page refresh, server-side routing does, thereby preventing RudderStack
      // from ever loading).
      // See here for more context:
      // GIF: https://github.com/benjaminhoffman/gatsby-plugin-segment/pull/19#issuecomment-559569483
      // TTI: https://github.com/GoogleChrome/lighthouse/blob/master/docs/scoring.md#performance
      // Problem/solution: https://marketingexamples.com/seo/performance
      delayLoad: false,
      // number (default to 1000); time to wait after scroll or route change
      // To be used when `delayLoad` is set to `true`
      delayLoadTime: 1000,
      // Whether to completely skip calling `analytics.load()`.
      // ADVANCED FEATURE: only use if you are calling `analytics.load()` manually
      // elsewhere in your code or are using a library
      // that will call it for you.
      // Useful for only loading the tracking script once a user has opted in to being tracked, for example.
      // *Another use case is if you want to add callbacks to the methods at load time.
      manualLoad: false,
      // Can be used to override where the SDK should be loaded from. This is useful
      // if you want to serve the SDK from a custom domain other than RudderStack to tackle ad-blockers
      // By default, the plugin will use the latest JS SDK from RudderStack's CDN
      sdkURL: `https://subdomain.yourdomain.com/v1.1/rudder-analytics.min.js`,
      // string ('async' or 'defer'); whether to load the RudderStack SDK async or defer. Anything else
      // will load normally.
      // 'async' will load the SDK as <script async></script>
      // 'defer' will load the SDK as <script defer></script>
      loadType: 'default',
      // Options to the `load` API
      // Note: The above `controlPlaneUrl` overrides the `configUrl` field in this object
      // You can find the loadOptions for JS SDK here:
      // https://www.rudderstack.com/docs/sources/event-streams/sdks/rudderstack-javascript-sdk/load-js-sdk/#loading-options
      loadOptions: {
        ...
      }
    }
  }
];

Step 2: Identify your users with the `identify()` method:

The identify() method allows you to link users and their actions to a specific user ID.

A sample example of how the identify() method works in Gatsby is as shown:

class CallToAction extends React.Component {
    _handleCallToAction() {
        window.rudderanalytics.identify(
          "12345", {
            email: "name@domain.com"
          }, {
            page: {
              path: "/post",
              referrer: "internal",
              search: "",
              title: "Post Page",
              url: "",
            },
          }
        }
        render() {
            return (
                <Link onClick={ this._handleCallToAction } to="/write-post">Write a Post</Link>
              )
        }
}

In the above example, information such as the user ID, email along with contextual information such as IP address, anonymousId, etc. will be captured.

There is no need to call identify() for anonymous visitors to your website. Such visitors are automatically assigned an anonymousId.

Step 3: Track your users’ actions with the `track()` method

The track() method allows you to track any actions that your users might perform. A sample example of how the track() method works is as shown:

window.rudderanalytics.track(
  "test track event GA3",
  {
    revenue: 30,
    currency: "USD",
    user_actual_id: 12345,
  },
  () => {
    console.log("in track call")
  }
)

In the above example, the method tracks the event ‘test track event GA3’, and information such as the revenue, currency, and anonymousId. You can use this method to track various other success metrics for your website, such as user signups, item purchases, article bookmarks, and much more.

To override contextual information, for ex: anonymizing IP and other contextual fields like page properties, the following template can be used. Similarly, you can override the auto-generated anonymousId with the provided ID. For this, follow the snippet below:

window.rudderanalytics.track(
  "test track event GA3",
  {
    revenue: 30,
    currency: "USD",
    user_actual_id: 12345,
  },
  () => {
    console.log("in track call")
  }
)

Tracking pageviews

If you want to track pageviews automatically, set trackPage to true in your gatsby-config.js file. What we mean by automatically is that whenever there is a route change, RudderStack leverages Gatsby’s onRouteUpdate API in the gatsby-browser.js file (link) to invoke window.rudderanalytics.page() on each route change.

However, if you want to pass in properties along with the pageview call (it is common to see some users pass in some user or account data with each page call), then you’ll have to set trackPage: false and call it yourself in your gatsby-browser.js file, as shown below:

// gatsby-browser.js
exports.onRouteUpdate = () => {
  window.rudderanalytics && window.rudderanalytics.page()
}

You’ve now successfully installed rudder-analytics.js tracking. You can enable and use any event destination to send your event data via RudderStack.

Step 4: Check ready state

There are cases when you may want to tap into the features provided by end destination SDKs to enhance tracking and other functionalities.

RudderStack’s JavaScript SDK exposes a ready API with a callback parameter, that fires when the SDK is done initializing itself and other third-party native SDK destinations. For example:

window.rudderanalytics.ready(() => {
  console.log("we are all set!!!")
})

For detailed technical documentation and troubleshooting guide on the RudderStack’s JavaScript SDK, see JavaScript SDK.

Using the Query string API

RudderStack’s Query string API allows you to trigger track, identify calls using query parameters.

If you pass the following parameters in the URL, then it will trigger the corresponding SDK API call. For example:

http://hostname.com/?ajs_uid=12345&ajs_event=test%20event&ajs_aid=abcde&ajs_prop_testProp=prop1&ajs_trait_name=Firstname+Lastname

For the above URL, the below SDK calls will be triggered:

rudderanalytics.identify("12345", { name: "Firstname Lastname" })
rudderanalytics.track("test event", { testProp: "prop1" })
rudderanalytics.setAnonymousId("abcde")

You may use the below parameters as a querystring parameter and trigger the corresponding call:

Parameter	Description
`ajs_uid`	Makes a `rudderanalytics.identify()` call with userId having the value of the parameter value.
`ajs_aid`	Makes a `rudderanalytics.setAnonymousId()` call with anonymousId having the value of the parameter value.
`ajs_event`	Makes a `rudderanalytics.track()` call with event name as parameter value.
`ajs_prop_`	If `ajs_event` is passed as querystring, value of this parameter will populate the properties of the corresponding event in the track call.
`ajs_trait_`	If `ajs_uid` is provided as querystring, value of this parameter will populate the traits of the identify call made.

Adding callbacks to standard methods

You can also define callbacks to the common methods of the rudderanalytics object.

Currently, the functionality is supported for syncPixel method which is called in the SDK when making sync calls in integrations for relevant destinations. For example, you can load rudderanalytics with callbacks in your Gatsby browser file as shown below:

window.rudderanalytics.syncPixelCallback = obj => {
  window.rudderanalytics.track(
    "sync lotame",
    { destination: obj.destination },
    { integrations: { All: false, S3: true } }
  )
}

In order for this to work well, you must also set the manualLoad option to true in your Gatsby configuration, and load RudderStack in the browser’s Gatsby file manually.

In the above example, we defined a syncPixelCallback on the rudderanalytics object before the call to load the SDK. This will lead to calling of this registered callback with the parameter {destination: <destination_name>} whenever a sync call is made from the SDK for relevant integrations like Lotame.

The callback can be supplied in options parameter like below as well:

// Define the callbacks directly on the load method like:
rudderanalytics.load(YOUR_WRITE_KEY, DATA_PLANE_URL, {
  clientSuppliedCallbacks: {
    syncPixelCallback: () => {
      console.log("sync done!")
    },
  },
})

License

This plugin is released under the AGPLv3 License.

Was this page helpful?

Glad to hear it! Please tell us how we can improve.

Sorry to hear that. Please tell us how we can improve.

Questions? Contact us by email or on Slack